Abstract

The Asian–Pacific Oscillation (APO) is a dominant teleconnection pattern linking the climate anomalies over Asia, the North Pacific, and other regions including North America. The National Centers for Environmental Prediction (NCEP) Climate Forecast System version 2 (CFSv2) successfully simulates many summer-mean features of the upper-tropospheric temperature, the South Asian high, the westerly and easterly jet streams, and the regional monsoons over Asia and Africa. It also well simulates the interannual variability of the APO and associated anomalies in atmospheric circulation, precipitation, surface air temperature (SAT), and sea surface temperature (SST). Associated with a positive APO are a strengthened South Asian high; a weakened extratropical upper-tropospheric westerly jet stream over North America; strengthened subtropical anticyclones over the Northern Hemisphere oceans; and strengthened monsoons over North Africa, India, and East Asia. Meanwhile, increased precipitation is found over tropical North Africa, South Asia, northern China, and tropical South America; decreased precipitation is seen over subtropical North Africa, the Middle East, central Asia, southern China, Japan, and extratropical North America. Low SAT occurs in North Africa, India, and tropical South America and high SAT appears in extratropical Eurasia and North America. SST increases in the extratropical Pacific and the North Atlantic but decreases in the tropical Pacific. The summer APO and many of the associated climate anomalies can be predicted by the NCEP CFSv2 by up to 5 months in advance. However, the CFSv2 skill of predicting the SAT in the East Asian monsoon region is low.

1. Introduction

A teleconnection pattern that links the climate over Asia and the North Pacific Ocean during summer has been identified recently. The pattern, named the Asian–Pacific Oscillation (APO) (Zhao et al. 2007), is measured by the zonal difference in upper-tropospheric temperature between Asia and the North Pacific. A positive (negative) APO phase corresponds to higher (lower) upper-tropospheric temperature over Asia and lower (higher) upper-tropospheric temperature over the North Pacific. That is, a positive (negative) APO indicates a strengthened (weakened) thermal contrast between Asia and the North Pacific (Zhao et al. 2012).

The APO measures the variability of many atmospheric circulation systems over the Northern Hemisphere and sea surface temperature (SST) in the tropical–extratropical North Pacific during summer. Associated with a positive APO are a stronger upper-tropospheric high over South Asia, a stronger lower-tropospheric low over Asia, stronger lower-tropospheric subtropical anticyclones over the North Pacific and the North Atlantic, and a westward stretch of the anticyclone from the North Atlantic to North America (Zhao et al. 2007, 2012). Accordingly, the summer monsoon precipitation over Northern Hemisphere lands mostly increases, with decreased precipitation over southern China and North America (Zhao et al. 2012). Significant anomalies are also found in land surface air temperature (SAT) and SST, with a warm SST–ridge structure over the northern oceans (Zhou et al. 2010; Zhao et al. 2012). Besides the above features that vary on interannual time scales, the APO is also associated with changes in atmospheric circulation and precipitation over Asia and North America on interdecadal time scales (Zhao et al. 2011).

Previous studies have also attempted to understand the APO and associated climate anomalies using climate models. Recently, Zhao et al. (2010) conducted a series of experiments using the National Center for Atmospheric Research (NCAR) Community Climate System Model version 3 (CCSM3) and the Community Atmospheric Model version 3 (CAM3). It was found that many features of the APO and related dynamical structure during summer could be captured by general circulation models. Nan et al. (2009) and Zhao et al. (2010) compared the relationship between APO and Pacific SST in the CCSM3 with observations and examined the impact of elevated heating over Asia on ocean–atmosphere interaction over the tropical–extratropical North Pacific. Moreover, using the coupled climate system model Flexible Global Ocean–Atmosphere–Land System Model gridpoint version 1.0 (FGOALS-gl.0) developed by the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics, the Institute of Atmospheric Physics of the Chinese Academy of Sciences, Chen et al. (2013) simulated the summer APO index and associated climate anomalies over the twentieth century, and Man and Zhou (2011) simulated a longer change in the summer APO index over the past millennium. The modeled result largely resembled the reconstruction of APO index. In addition, Zhou and Zhao (2010a, 2012) used the CCSM3 to investigate the relationship between the APO and associated climate anomalies over East Asia in the Middle Holocene.

The National Centers for Environmental Prediction (NCEP) Climate Forecast System (CFS) is a state-of-the-art climate forecast system. Since it became operational at NCEP in August 2004, the model has provided monthly and seasonal climate predictions over the world. It has been proved that the CFS is competitive in forecasting extratropical SAT and precipitation with some major statistical methods (Van Oldenborgh et al. 2003; Saha et al. 2006). The model demonstrates skills in simulating and predicting the variability of El Niño–Southern Oscillation (ENSO) (Wang et al. 2005) and the tropical Atlantic SST (Hu and Huang 2007; Misra et al. 2009). It also has high skill in predicting the precipitation over South Africa (Wang et al. 2010) and the United States (Goddard et al. 2006; Higgins et al. 2008; Yoon et al. 2012) and the upper-tropospheric circulation over the Northern Hemisphere (Lee et al. 2011). Moreover, the CFS successfully simulates the major climatological features and the subseasonal-to-interannual variations of the Asian summer monsoon (Yang et al. 2008a,b; Achuthavarier and Krishnamurthy 2010) and is capable of predicting the most dominant modes of the Asian and Indo-Pacific summer precipitation by several months in advance (Liang et al. 2009). Thus, CFS products have been extensively applied to the regional climate prediction in many Asian countries. Recently, CFS version 2 (CFSv2) replaced the early version of the forecast system and became operational at NCEP in March 2011 (Saha et al. 2012, manuscript submitted to J. Climate). Compared to its earlier version, the NCEP CFSv2 has overall higher skills in predicting the world’s climate (see the proceedings of the 2012 CFSv2 Evaluation Workshop, available online at http://www.joss.ucar.edu/events/2012/cfsv2/index.html), including the Asian monsoon (e.g., Jiang et al. 2013a,b; Liu et al. 2013).

Nevertheless, the NCEP CFS also faces many difficulties in predicting various regional features of the Asian monsoon climate. For example, a recent study shows that the model does not capture the observed negative correlation between ENSO and the Indian summer monsoon realistically (Achuthavarier et al. 2012). Jiang et al. (2013a) and Liu et al. (2013) have also demonstrated that the CFSv2 overestimates the convection over the equatorial Indian Ocean and underestimates the intensity of the subtropical western Pacific high, although it improves the cold bias that existed in the early version of the model (see Yang et al. 2008b). Then, does the CFSv2 capture the variability of the large-scale APO pattern and the associated anomalies of atmospheric circulation, precipitation, and SST? If yes, can the features be predicted by the CFSv2 with reasonable skills and how do the skills depend on lead time? Given the importance of the APO in linking the climate anomalies over a large portion of the Northern Hemisphere and the importance of the CFSv2 as a state-of-the-art climate forecast system, we carry out this analysis to address the above questions.

The rest of this paper is organized as follows. We describe the main features of the model, data, and analysis methods applied in this paper in section 2, and discuss the climatological features of summer atmospheric circulation and precipitation over the Asian–Pacific region, the APO pattern, and APO-related atmospheric circulation, rainfall, SAT, and SST anomalies simulated by the CFSv2 in section 3. In section 4, we discuss the prediction of the APO and associated climate anomalies by the CFSv2. Finally, a summary of the study and further discussion are provided in section 5.

2. Model, hindcast, and observational data

The atmospheric component of CFSv2 is the NCEP Global Forecast System model for operational weather forecast (Moorthi et al. 2001). The oceanic component is upgraded from the National Oceanic and Atmospheric Administration (NOAA) Geophysical Fluid Dynamics Laboratory (GFDL) Modular Ocean Model (MOM) version 3.0 to MOM version 4.0 (Griffies et al. 2004). The land surface model is the four-layer NCEP, the Oregon State University (OSU), the Air Force, and the Hydrologic Research Laboratory (Noah) land model including the global land data assimilation (Koren et al. 1999; Chen and Dudhia 2001; Ek et al. 2003). A three-layer interactive global sea ice model is also introduced to the CFSv2 [see Saha et al. (2012, manuscript submitted to J. Climate) for details].

We analyze the output from the CFSv2 retrospective predictions that cover all 12 calendar months from 1982 to 2010. These hindcast runs, each of which is for 9-month integration, are initiated by using perturbed real-time oceanic and atmospheric initial conditions (ICs) from the Climate Forecast System Reanalysis (CFSR; Saha et al. 2010). For each year of 1982–2010, the runs begin from 1 January and are initiated with an interval of 5 days and run from all four cycles of that day. Moreover, the model repeats 4 times using initial data at 0000, 0600, 1200, and 1800 UTC. Thus, there are 24 members for each month. Unless specified, the values of CFSv2 presented in this paper are the ensemble means of 24 members, and the climatological means are for 1982–2010. For a convenient comparison with observations, the CFSv2 hindcast data with a horizontal resolution of 1° in both longitude and latitude are interpolated on the grid points of the observed data by using a bilinear interpolation method.

The observational datasets used in model verification include the monthly precipitation from the Global Precipitation Climatology Project (GPCP) (Adler et al. 2003). They also include the monthly-mean winds, geopotential height, and temperatures from the NCEP–Department of Energy (DOE) Reanalysis 2 (Kanamitsu et al. 2002).

Correlation and regression analyses are used to examine relationships between two variables. Empirical orthogonal function (EOF) analysis with latitudinal weighting is also carried out. The statistical significance of a relationship is assessed using the Student’s t test, and the 95% confidence level is used unless otherwise stated. In this study, the summer season refers to the months from June to August (JJA).

3. Simulation of APO and associated climate anomalies

a. Climatological features of atmospheric circulation over Eurasia and the North Pacific

In this section, we analyze the CFSv2 simulations for JJA climate using the ICs from 1 June to 20 June. We first examine the features of upper-tropospheric temperature because the APO is defined based on the departure of upper-tropospheric temperature (T′), which is obtained by removing the zonal mean temperature () from the total temperature (T). Figure 1a shows the climatological mean (1982–2010 average) of simulated summer upper-tropospheric (300–200 hPa) T′. Positive values appear in the middle to lower latitudes of Eurasia, with a central value of 5°C over Asia, and negative values occur over the central-eastern North Pacific, with a central value of −3°C. These features in CFSv2 are consistent with those shown in the NCEP–DOE reanalysis (Fig. 1b).

Fig. 1.

Climatology of JJA mean 300–200-hPa T′ (°C) during 1982–2010 from (a) CFSv2 simulation (0-month lead) and (b) NCEP–DOE reanalysis 2.

Fig. 1.

Climatology of JJA mean 300–200-hPa T′ (°C) during 1982–2010 from (a) CFSv2 simulation (0-month lead) and (b) NCEP–DOE reanalysis 2.

Figure 2a shows the climatology of simulated summer-mean 850-hPa winds. The CFSv2 captures many major features of the Asian summer monsoon including the cross-equatorial flows off Somalia and over 105°–130°E and the lower-tropospheric southwesterly and southerly winds over tropical Asia and extratropical East Asia. At 500 hPa (Fig. 2a), the high pressure system over the subtropical western North Pacific and the deep trough over extratropical East Asia are also simulated reasonably well. At 200 hPa (Fig. 2b), a large-scale anticyclone appears over the middle to lower latitudes of North Africa and Eurasia, with a center over the Tibetan Plateau (i.e., the South Asian high). The westerly (easterly) jet stream to the north (south) of the South Asian high prevails over the middle to higher latitudes of Eurasia and the North Pacific (the tropics from Africa to the western Pacific). Compared to observations (Figs. 2c,d), the locations of the major large-scale circulation systems in the troposphere are simulated reasonably well, although the intensity of the South Asian high and the high pressure system over the western North Pacific is underestimated by the CFSv2.

Fig. 2.

Climatology (1982–2010) of JJA CFSv2 simulation (0-month lead) for (a) 850-hPa winds (m s−1; vectors) and 500-hPa geopotential height (m; shaded) and (b) 200-hPa winds (m s−1; vectors) and geopotential height (m; shaded). (c),(d) As in (a),(b), but for the NCEP–DOE reanalysis 2. The green thick solid lines indicate the topographic contour of 1500 m.

Fig. 2.

Climatology (1982–2010) of JJA CFSv2 simulation (0-month lead) for (a) 850-hPa winds (m s−1; vectors) and 500-hPa geopotential height (m; shaded) and (b) 200-hPa winds (m s−1; vectors) and geopotential height (m; shaded). (c),(d) As in (a),(b), but for the NCEP–DOE reanalysis 2. The green thick solid lines indicate the topographic contour of 1500 m.

b. Climatological features of monsoon precipitation

Figure 3a shows the climatology of simulated JJA-mean precipitation in the CFSv2. A rain belt appears over the Asian monsoon region and the tropical North Pacific, with centers over western India, the Bay of Bengal, and the eastern South China Sea, respectively. Moreover, rainfall above 6 mm day−1 occurs from eastern China to southern Japan, representing the mei-yu in eastern China, the baiu in Japan, and the changma in Korea. Compared to observations (Fig. 3b), the CFSv2 overestimates the precipitation over the Asian and Pacific monsoon regions. Figure 3c, which shows the difference in JJA precipitation between CFSv2 and GPCP, further indicates positive differences over the Indian Ocean, the South China Sea, and the western Pacific, with scattered negative differences over the South Asian monsoon region. Precipitation differences are relatively small over East Asia, mostly between −2 and 2 mm day−1. In spite of these discrepancies, the simulated precipitation pattern is overall similar to that observed. In particular, the mei-yu, baiu, and changma rainbands are also simulated reasonably well. These results show that the CFSv2 captures the major features of precipitation over the Asian–Pacific monsoon regions.

Fig. 3.

(a) Climatology (1982–2010) of JJA mean precipitation (mm day−1) from CFSv2 simulation (0-month lead). (b) As in (a), but for the GPCP precipitation. (c) Difference of (a) and (b).

Fig. 3.

(a) Climatology (1982–2010) of JJA mean precipitation (mm day−1) from CFSv2 simulation (0-month lead). (b) As in (a), but for the GPCP precipitation. (c) Difference of (a) and (b).

c. Asian–Pacific Oscillation and associated climate anomalies

1) Asian–Pacific oscillation

The APO reviewed in the introduction may also be defined as the most dominant mode of the upper-tropospheric T′ over the Northern Hemisphere. Following Zhao et al. (2010, 2012), we perform an EOF analysis on the anomaly of simulated summer 300–200-hPa T′ for the period of 1982–2010. The first EOF mode (hereafter EOF1) accounts for 50% of the total variation and exhibits an out-of-phase relationship between the North African–Eurasian sector and the North Pacific, with positive values exceeding 0.02 over lands and negative values below −0.02 over the North Pacific (Fig. 4a). This EOF1 pattern is consistent with the result from the NCEP–DOE reanalysis (Fig. 4b) and similar to that from the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Zhao et al. 2007).

Fig. 4.

(a) EOF1 (×0.01) of the anomalies of JJA upper-tropospheric (300–200-hPa) mean T′ from the CFSv2 simulation (0-month lead, positive values shaded). (b) As in (a), but for the NCEP–DOE reanalysis 2.

Fig. 4.

(a) EOF1 (×0.01) of the anomalies of JJA upper-tropospheric (300–200-hPa) mean T′ from the CFSv2 simulation (0-month lead, positive values shaded). (b) As in (a), but for the NCEP–DOE reanalysis 2.

Following Zhao et al. (2007), we also analyze the zonal difference in extratropical upper-tropospheric (300–200 hPa) temperature between Asia (15°–50°N, 60°–120°E) and the North Pacific (15°–50°N, 180°–120°W)—the so-called APO index. Apparently, this difference calculated from T is consistent with that from T′. Figure 5 shows the time series of the simulated standardized APO index. For the entire study period, the simulated summer APO index exhibits an apparent interannual variation, similar to the observed (Fig. 5). A correlation coefficient of 0.65 is obtained between the simulated APO and the observed APO for 1982–2010, significantly exceeding the 99.9% confidence level.

Fig. 5.

CFSv2 simulated (0-month lead; blue) and observed (red) standardized summer APO indices during 1982–2010.

Fig. 5.

CFSv2 simulated (0-month lead; blue) and observed (red) standardized summer APO indices during 1982–2010.

From the above analyses, both the summer APO pattern and its interannual variability simulated by the CFSv2 are highly similar to those observed. In the following sections, we analyze the APO-related climate anomalies simulated and predicted by the model.

2) APO-related anomalies of atmospheric circulation, precipitation, and surface temperature

Figure 6a shows the regression of summer 200-hPa winds against the APO index in the CFSv2. A large-scale anomalous anticyclonic circulation appears over Eurasia and the western North Pacific, with centers near western Asia, northeast Asia, and the North Pacific, indicating a strengthened South Asian high. Westerly (easterly) anomalies prevail to the north (south) of the anomalous circulation centers over extratropical Eurasia (the tropics from East Asia to North Africa), indicating a strengthened westerly (easterly) jet stream. Meanwhile, easterly anomalies prevail over extratropical North America, indicating a weakened local westerly jet stream. At 850 hPa (Fig. 6b), two centers of anomalous anticyclonic circulation appear over the subtropical North Pacific and Atlantic oceans, near 35°N, 170°W and 40°N, 50°W, respectively, indicating stronger subtropical anticyclones over these oceans. Anomalous cyclonic circulations are found over North Africa and the Arabian Sea. Westerly anomalies generally prevail over the tropics from North Africa to southern India and southerly or southeasterly anomalies prevail over the East Asian monsoon region. This result indicates strengthened southwesterly monsoon flow over these regions when the APO index is higher. In the meantime, the anomalous anticyclonic circulation over the North Atlantic stretches westward into extratropical North America, implying a weakened low pressure system over extratropical North America. In addition, large-scale easterly anomalies prevail over the tropical Pacific. These atmospheric circulation anomalies in the CFSv2 are similar to those obtained from the reanalysis (Zhao et al. 2007, 2012).

Fig. 6.

(a) Regression of JJA 200-hPa winds (m s−1) against the summer APO index from the CFSv2. (b) As in (a), but for 850-hPa winds (the green thick solid lines indicate the topographic contours of 1500 m). (c) As in (a), but for summer mean precipitation (mm day−1). (d) As in (a), but for surface air temperature (°C) over land and SST (°C) over oceans. Values significantly above the 95% (90%) confidence level are shaded in (a) and (b) [(c) and (d)].

Fig. 6.

(a) Regression of JJA 200-hPa winds (m s−1) against the summer APO index from the CFSv2. (b) As in (a), but for 850-hPa winds (the green thick solid lines indicate the topographic contours of 1500 m). (c) As in (a), but for summer mean precipitation (mm day−1). (d) As in (a), but for surface air temperature (°C) over land and SST (°C) over oceans. Values significantly above the 95% (90%) confidence level are shaded in (a) and (b) [(c) and (d)].

The observational diagnosis has shown that associated with the large-scale atmospheric circulation anomalies are enhanced precipitation over the Northern Hemisphere summer monsoon regions and reduced precipitation over their adjacent arid regions to the west and north of the monsoon domain such as subtropical North Africa, the Middle East, central Asia, and extratropical North America (Zhao et al. 2012). Figure 6c shows the regression of summer precipitation against the APO index in the CFSv2. Significant positive anomalies appear over tropical North Africa, the Asian monsoon region (including India, the Indo-China Peninsula, and north China), the Mexican monsoon region, and the tropics of South America. This result suggests enhanced summer monsoon rainfall over the lands of all major Northern Hemisphere monsoon regions (North Africa, South Asia, East Asia, Mexico, and South America) when the simulated APO index is higher. Meanwhile, negative precipitation anomalies occur over subtropical North Africa, the Middle East, central Asia, southern China, Japan, and extratropical North America. This simulated relationship between the APO index and the Northern Hemisphere precipitation is overall consistent with the observed results (Zhao et al. 2007, 2012).

Figure 6d shows the regression of surface temperature against the summer APO index in the CFSv2. Both SAT over lands and SST are used in the computation. Corresponding to a positive APO phase, negative temperature anomalies appear in North Africa, India, and tropical South America, while positive temperature anomalies cover most of extratropical Eurasia and North America. These positive and negative anomalies of SAT are generally consistent with those from observations (Zhao et al. 2012). Over oceans, positive temperature anomalies appear in the midlatitudes of the North Pacific and the South Pacific, the tropical Atlantic, and the midlatitudes of the North Atlantic, while large-scale negative temperature anomalies occur in the tropical central-eastern Pacific, indicating a La Niña condition. These anomalies associated with the APO in the CFSv2 are also in agreement with observations (Zhao et al. 2012).

In brief, the anomalies of simulated atmospheric circulation, precipitation, SAT, and SST associated with the summer APO in the CFSv2 are highly consistent with observations. Such a good agreement is partly due to the use of real-time atmospheric and oceanic initial conditions in the model. This consistency demonstrates the capability of CFSv2 in simulating the major climate anomalies over the Northern Hemisphere.

4. Predictions of the APO and associated climate anomalies

In the hindcast of 9-month integration, the prediction of the 7-month lead is the longest lead available for analyzing seasonal means. That is, the 7-month lead using the ICs of November is the longest lead for predicting the summer APO index and associated climate anomalies. Here, we denote the 1-month lead as LM1, 2-month lead as LM2, …, and 7-month lead as LM7.

Figure 7a shows the EOF1 pattern of summer upper-tropospheric T′ for 1-month lead (LM1) using May ICs. This dominant mode accounts for 50% of the total variation, showing positive values over North Africa and Eurasia but negative values over the North Pacific, similar to the EOF1 pattern of LM0 (Fig. 4a). Moreover, the EOF1 modes from LM2 to LM7 (Figs. 7b–g) also exhibit features that are similar to those of the LM0 pattern. Figure 8 further shows the correlation between the observed summer APO index and the predicted indices of different leads. The correlation coefficient decreases gradually from LM1 to LM7 but exceeds 0.40 in LM1, LM2, and LM3. The coefficient in LM1 is 0.63, significantly exceeding the 99.9% confidence level, and the coefficient in LM7 is 0.39, significantly exceeding the 95% confidence level.

Fig. 7.

(a)–(g) As Fig. 4a, but for the predicted JJA EOF1 for the 1-month lead (LM1) to the 7-month lead (LM7), respectively. Positive values are shaded.

Fig. 7.

(a)–(g) As Fig. 4a, but for the predicted JJA EOF1 for the 1-month lead (LM1) to the 7-month lead (LM7), respectively. Positive values are shaded.

Fig. 8.

Coefficients of correlation between the JJA APO index predicted by the CFSv2 of various leads and the observed JJA APO index during 1982–2010.

Fig. 8.

Coefficients of correlation between the JJA APO index predicted by the CFSv2 of various leads and the observed JJA APO index during 1982–2010.

Figure 9 shows the regression of LM0 summer 200-hPa winds against the APO index in LM1, LM3, and LM5. For the LM1 APO index (Fig. 9a), the summer anomalous anticyclonic circulations over the middle to lower latitudes from North Africa to the North Pacific are well forecasted, with westerly (easterly) anomalies prevailing over extratropical Asia (tropical North Africa and the tropical North Pacific). Meanwhile, easterly anomalies appear over extratropical North America. These anomalous features are also forecasted in LM3 and LM5 (see Figs. 9b,c).

Fig. 9.

Regression of LM0 JJA 200-hPa winds (m s−1) against the summer APO index predicted by the CFSv2 for 1982–2010, for (a) 1-, (b) 3-, and (c) 5-month leads. Shaded values are significantly above the 95% confidence level.

Fig. 9.

Regression of LM0 JJA 200-hPa winds (m s−1) against the summer APO index predicted by the CFSv2 for 1982–2010, for (a) 1-, (b) 3-, and (c) 5-month leads. Shaded values are significantly above the 95% confidence level.

Figure 10a shows the regression of LM0 summer 850-hPa winds against the APO index in LM1. Anomalous anticyclonic circulations appear over the extratropical South and North Pacific Oceans and the North Atlantic, with centers near 35°S, 150°W; 40°N, 175°W; and 45°N, 60°W, respectively. On the other hand, anomalous cyclonic circulations occur over North Africa, the Arabian Sea, and the Bay of Bengal. Westerly anomalies prevail over North Africa and southerly or southeasterly wind anomalies are dominant over northern China. Large-scale easterly anomalies are found over the tropical Pacific. Moreover, the anomalous anticyclonic circulation over the North Atlantic expands westward into extratropical North America. Comparison of the features in Fig. 10a with the results in Fig. 10b (for LM3) and Fig. 10c (for LM5) indicates that the southwesterly or southerly anomalies and anomalous anticyclonic circulations over the Northern Hemisphere oceans in LM1 are also predicted in LM3 and LM5.

Fig. 10.

As in Fig. 9, but for 850-hPa winds.

Fig. 10.

As in Fig. 9, but for 850-hPa winds.

It is evident that these predicted summer APO and atmospheric circulation anomalies by the CFSv2 are similar to those in observations and CFSv2 LM0. Therefore, the above results clearly show that the summer APO index and associated atmospheric circulation anomalies can be well predicted by the CFSv2 by up to 5 months in advance and the prediction using May ICs has the highest skill.

Corresponding to the predicted APO index and atmospheric circulation anomalies, the associated precipitation and surface temperature anomalies are also well predicted by the model by 5 months in advance. Figure 11 shows the regression of LM0 JJA rainfall against the APO index in LM1, LM3, and LM5. For LM1 (Fig. 11a), positive precipitation anomalies appear over tropical North Africa, India, the Indo-China Peninsula, north China, and tropical South America. Negative precipitation anomalies are seen over subtropical North Africa, the Middle East, central Asia, and extratropical North America. Similar anomalous features in precipitation over the North African–Asian monsoon regions, tropical South America, and extratropical North America are generally seen in the results of LM3 (Fig. 11b) and LM5 (Fig. 11c). For SAT (Fig. 12), the cooling over Africa, South Asia, and tropical South America and the warming over extratropical North America are predicted in LM1, LM3, and LM5. However, the increase in temperature over the high latitudes of East Asia can be predicted only in LM1 but no large-scale significant positive anomalies over East Asia can be seen in LM3 and LM5, meaning that there is a lower skill of the CFSv2 in predicting the summer East Asian SAT anomalies associated with APO.

Fig. 11.

As in Fig. 9, but for JJA mean precipitation (mm day−1). Shaded values are significantly above the 90% confidence level.

Fig. 11.

As in Fig. 9, but for JJA mean precipitation (mm day−1). Shaded values are significantly above the 90% confidence level.

Fig. 12.

As in Fig. 9, but for surface air temperature (°C) over land and SST (°C) over oceans. Shaded values are significantly above the 90% confidence level.

Fig. 12.

As in Fig. 9, but for surface air temperature (°C) over land and SST (°C) over oceans. Shaded values are significantly above the 90% confidence level.

Compared to the SAT over lands, the CFSv2 shows a higher skill in predicting the SST anomalies associated with APO. For the LM1 APO index (Fig. 12a), associated with a positive APO phase are significant positive SST anomalies over the central-western extratropics of the South Pacific, the North Pacific, and the North Atlantic and large-scale significant negative SST anomalies over the tropical central-eastern Pacific. The positive SST anomalies are well coupled with local lower-tropospheric anomalous anticyclones, indicating a warm SST–ridge relationship. Also, the negative SST anomalies match local lower-tropospheric easterly anomalies. These features represent a La Niña condition. Meanwhile, the SST anomalies in the tropical Pacific and the extratropical North Atlantic are successfully predicted in LM3 (Fig. 12b) and LM5 (Fig. 12c), which indicates that the CFSv2 has a higher skill in predicting SST anomalies compared to predicting SAT anomalies over lands.

5. Summary and discussion

The products of the NCEP operational Climate Forecast System version 2 (CFSv2) are now becoming an important source of information for regional climate predictions in many Asian countries where monsoon climate dominates. The summer Asian–Pacific Oscillation is a major mode of climate variations over the subtropics, especially in the Asian–Pacific sector, and is also a good index to measure monsoon precipitation, SAT, and SST anomalies (Zhao et al. 2012). In this study, we have used the NCEP CFSv2 retrospective ensemble hindcast to understand the predictability of the summer APO pattern and associated atmospheric circulation, precipitation, and surface temperature anomalies.

The CFSv2 successfully captures many major climatological features of the atmospheric circulation systems over Eurasia and the North Pacific including the upper-tropospheric South Asian high and jet streams, lower-tropospheric southwesterly or southerly monsoon flow, and precipitation over Africa and Asia. Compared to the early version of the CFS (Yang et al. 2008b), the CFSv2 improves the ability of simulating these atmospheric circulation systems, although it still produces weaker than observed large-scale atmospheric circulation over some regions (also see Jiang et al. 2013a). The CFSv2 well simulates the summer APO pattern between Eurasia and the North Pacific as well as its interannual variability observed during 1982–2010.

For the upper troposphere, the CFSv2 captures the strengthened South Asian high, extratropical westerly jet stream over Eurasia, and tropical easterly jet stream over Asia and North Africa and the weakened extratropical westerly jet stream over North America during a positive APO phase. For the lower troposphere, the strengthened subtropical anticyclones over the North Pacific and the North Atlantic, the westward stretch of the anticyclone over the North Atlantic into extratropical North America, and the strengthened southwesterly monsoon circulation over North Africa, India, and East Asia are also well simulated. More precipitation appears over tropical North Africa, South Asia, north China, and South America, while less precipitation occurs over subtropical North Africa, the Middle East, central Asia, southern China, Japan, and extratropical North America. Meanwhile, low SAT is found in North Africa, India, and tropical South America, and high SAT is seen in extratropical Eurasia and North America. SST increases in the extratropical South Pacific and North Pacific, as well as the extratropical North Atlantic, but decreases in the tropical central-eastern Pacific. These anomalies of atmospheric circulation, precipitation, SAT, and SST simulated by the CFSv2 are similar to those observed (Zhao et al. 2012), demonstrating the capability of the CFSv2 in simulating the summer APO and associated climate anomalies.

The CFSv2 can predict the summer APO pattern reasonably well by 7 months in advance, with a coefficient of 0.63 for the correlation between observation and the prediction of 1-month lead during 1982–2010. The major features of large-scale atmospheric circulation, precipitation, SAT, and SST associated with the APO anomaly can also be reasonably predicted by 5 months in advance. These features include the South Asian high, the African–Asian monsoon circulation, and the precipitation over North Africa, Eurasia, and extratropical North America and South America, as well as the SSTs in the Pacific and the Atlantic.

Although the CFSv2 is capable of predicting the summer APO teleconnection and associated climate anomalies, it also shows a lower skill in forecasting the SAT anomalies in East Asian lands. This lower skill may be attributed to the complicated land–atmosphere interaction over the land regions and the drawback of the CFSv2 in describing the interaction. However, in spite of this shortcoming of the model, the CFSv2 is an important tool for predicting the climate anomalies over the subtropics that are related to APO.

Apparent features associated with the APO and its link to climate anomalies have also been observed in other seasons (Zhou and Zhao 2010b; Zou and Zhao 2011). It will also be interesting to further understand the capability of the NCEP CFSv2 in simulating and forecasting the APO and associated climate anomalies for the nonsummer seasons.

Acknowledgments

This work was sponsored by the Basic Research Fund of the Chinese Academy of Meteorological Sciences (Grants 2010Z001 and 2010Z003), the Special Project of the China Meteorological Administration (Grant GYHY200906017), the National Key Basic Research Project of China (2009CB421404), and the National Natural Science Foundation of China (40890052 and 40890053).

REFERENCES

REFERENCES
Achuthavarier
,
D.
, and
V.
Krishnamurthy
,
2010
:
Relation between intraseasonal and interannual variability of the South Asian monsoon in the National Centers for Environmental Prediction forecast systems
.
J. Geophys. Res.
,
115
,
D08104
,
doi:10.1029/2009JD012865
.
Achuthavarier
,
D.
,
V.
Krishnamurthy
,
B. P.
Kirtman
, and
B.
Huang
,
2012
:
Role of the Indian Ocean in the ENSO–Indian summer monsoon teleconnection in the NCEP Climate Forecast System
.
J. Climate
,
25
,
2490
2508
.
Adler
,
R. F.
, and
Coauthors
,
2003
:
The version 2 Global Precipitation Climatology Project (GPCP) Monthly Precipitation Analysis (1979–present)
.
J. Hydrometeor.
,
4
,
1147
1167
.
Chen
,
F.
, and
J.
Dudhia
,
2001
:
Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity
.
Mon. Wea. Rev.
,
129
,
569
585
.
Chen
,
X.
,
T.
Zhou
, and
L.
Zou
,
2013
:
Variation of the Asian-Pacific Oscillation simulated by LASG/IAP Climate System Model FGOALS_gl in boreal summer
.
Acta Meteor. Sin.
,
in press
.
Ek
,
M.
,
K.
Mitchell
,
Y.
Lin
,
E.
Rogers
,
P.
Grunmann
,
V.
Koren
,
G.
Gayno
, and
J. D.
Tarpley
,
2003
:
Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model
.
J. Geophys. Res.
,
108
,
8851
,
doi:10.1029/2002JD003296
.
Goddard
,
L.
,
A.
Kumar
,
M. P.
Hoerling
, and
A. G.
Barnston
,
2006
:
Diagnosis of anomalous winter temperatures over the eastern United States during 2002/03 El Niño
.
J. Climate
,
19
,
5624
5636
.
Griffies
,
S. M.
,
M. J.
Harrison
,
R. C.
Pacanowski
, and
A.
Rosati
,
2004
: A technical guide to MOM4. GFDL Ocean Group Tech. Rep. 5, 342 pp. [Available online at http://www.gfdl.noaa.gov/bibliography/related_files/smg0301.pdf.]
Higgins
,
R. W.
,
V. B. S.
Silva
,
V. E.
Kousky
, and
W.
Shi
,
2008
:
Comparison of daily precipitation statistics for the United States in observations and in the NCEP Climate Forecast System
.
J. Climate
,
21
,
5993
6014
.
Hu
,
Z.-Z.
, and
B.
Huang
,
2007
:
The predictive skill and the most predictable pattern in the tropical Atlantic: The effect of ENSO
.
Mon. Wea. Rev.
,
135
,
1786
1806
.
Jiang
,
X.
,
S.
Yang
,
Y.
Li
,
A.
Kumar
,
X.
Liu
,
Z.
Zuo
, and
B.
Jha
,
2013a
: Seasonal-to-interannual prediction of the Asian summer monsoon in the NCEP Climate Forecast System version 2. J. Climate, in press.
Jiang
,
X.
,
S.
Yang
,
Y.
Li
,
A.
Kumar
,
W.
Wang
, and
Z.
Gao
,
2013b
:
Dynamical prediction of the East Asian winter monsoon by the NCEP Climate Forecast System. J. Geophys. Res., 118, 1312–1328, doi:10.1002/jgrd.50193
.
Kanamitsu
,
M.
,
W.
Ebisuzaki
,
J.
Woollen
,
S.-K.
Yang
,
J. J.
Hnilo
,
M.
Fiorino
, and
G. L.
Potter
,
2002
:
NCEP–DOE AMIP-II Reanalysis (R-2)
.
Bull. Amer. Meteor. Soc.
,
83
,
1631
1643
.
Koren
,
V.
,
J. C.
Schaake
,
K. E.
Mitchell
,
Q. Y.
Duan
,
F.
Chen
, and
J.
Baker
,
1999
:
A parameterization of snowpack and frozen ground intended for NCEP weather and climate models
.
J. Geophys. Res.
,
104
,
19 569
19 585
.
Lee
,
J.-Y.
,
B.
Wang
,
Q.
Ding
,
K.-J.
Ha
,
J.-B.
Ahn
,
A.
Kumar
,
B.
Stern
, and
O.
Alves
,
2011
:
How predictable is the Northern Hemisphere summer upper-tropospheric circulation?
Climate Dyn.
,
37
,
1189
1203
.
Liang
,
J.
,
S.
Yang
,
Z.-Z.
Hu
,
B.
Huang
,
A.
Kumar
, and
Z.
Zhang
,
2009
:
Predictable patterns of the Asian and Indo-Pacific summer precipitation in the NCEP CFS
.
Climate Dyn.
,
32
,
989
1001
.
Liu
,
X.
,
S.
Yang
,
A.
Kumar
,
S.
Weaver
, and
X.
Jiang
,
2013
:
Diagnostics of subseasonal prediction biases of the Asian summer monsoon by the NCEP Climate Forecast System
.
Climate Dyn.
,
doi:10.1007/s00382-012-1553-3, in press
.
Man
,
W. M.
, and
T. J.
Zhou
,
2011
:
Forced response of atmospheric oscillations during the last millennium by a climate system model
.
Chin. Sci. Bull.
,
56
,
3042
3052
.
Misra
,
V.
,
S.
Chan
,
R.
Wu
, and
E.
Chassignet
,
2009
:
Air–sea interaction over the Atlantic warm pool in the NCEP CFS
.
Geophys. Res. Lett.
,
36
,
L15702
,
doi:10.1029/2009GL038737
.
Moorthi
,
S.
,
H.-L.
Pan
, and
P.
Caplan
,
2001
: Changes to the 2001 NCEP operational MRF/AVN global analysis/forecast system. NWS Tech. Procedures Bulletin 484, 14 pp. [Available online at http://www.nws.noaa.gov/om/tpb/484.htm.]
Nan
,
S.
,
P.
Zhao
,
S.
Yang
, and
J. M.
Chen
,
2009
:
Springtime tropospheric temperature over the Tibetan Plateau and evolutions of the tropical Pacific SST
.
J. Geophys. Res.
,
114
,
D10104
,
doi:10.1029/2008JD011559
.
Saha
,
S.
, and
Coauthors
,
2006
:
The NCEP Climate Forecast System
.
J. Climate
,
19
,
3483
3517
.
Saha
,
S.
, and
Coauthors
,
2010
:
The NCEP Climate Forecast System Reanalysis
.
Bull. Amer. Meteor. Soc.
,
90
,
1015
1057
.
Van Oldenborgh
,
G. J.
, and
Coauthors
,
2003
: Did the ECMWF seasonal forecast model outperform a statistical model over the last 15 years? ECMWF Tech. Memo. 418, 32 pp.
Wang
,
W.
,
S.
Saha
,
H.-L.
Pan
,
S.
Nadiga
, and
G.
White
,
2005
:
Simulation of ENSO in the new NCEP coupled forecast system model
.
Mon. Wea. Rev.
,
133
,
1574
1593
.
Wang
,
W.
,
M.
Chen
, and
A.
Kumar
,
2010
:
An assessment of the CFS real-time seasonal forecasts
.
Wea. Forecasting
,
25
,
950
969
.
Yang
,
S.
,
M.
Wen
, and
W.
Higgins
,
2008a
:
Subseasonal features of the Asian summer monsoon in the NCEP climate forecast system
.
Acta Oceanol. Sin.
,
27
,
88
103
.
Yang
,
S.
,
Z.
Zhang
,
V. E.
Kousky
,
R. W.
Higgins
,
S.-H.
Yoo
,
J.
Liang
, and
Y.
Fan
,
2008b
:
Simulations and seasonal prediction of the Asian summer monsoon in the NCEP Climate Forecast System
.
J. Climate
,
21
,
3755
3775
.
Yoon
,
J.-H.
,
K.
Mo
, and
E. F.
Wood
,
2012
:
Dynamic-model-based seasonal prediction of meteorological drought over the contiguous United States
.
J. Hydrometeor.
,
13
,
463
482
.
Zhao
,
P.
,
Y.
Zhu
, and
R.
Zhang
,
2007
:
An Asian–Pacific teleconnection in summer tropospheric temperature and associated Asian climate variability
.
Climate Dyn.
,
29
,
293
303
.
Zhao
,
P.
,
Z.
Cao
, and
J.
Chen
,
2010
:
A summer teleconnection pattern over the extratropical Northern Hemisphere and associated mechanisms
.
Climate Dyn.
,
35
,
523
534
.
Zhao
,
P.
,
S.
Yang
,
H.
Wang
, and
Q.
Zhang
,
2011
:
Interdecadal relationships between the Asian–Pacific Oscillation and summer climate anomalies over Asia, North Pacific, and North America during a recent 100 years
.
J. Climate
,
24
,
4793
4799
.
Zhao
,
P.
,
B.
Wang
, and
X. J.
Zhou
,
2012
:
Boreal summer continental monsoon rainfall and hydroclimate anomalies associated with the Asian–Pacific Oscillation
.
Climate Dyn.
,
39
,
1197
1207
,
doi:10.1007/s00382-012-1348-6
.
Zhou
,
B.
, and
P.
Zhao
,
2010a
:
Modeling variations of summer upper tropospheric temperature and associated climate over the Asian Pacific region during the mid-Holocene
.
J. Geophys. Res.
,
115
,
D20109
,
doi:10.1029/2010JD014029
.
Zhou
,
B.
, and
P.
Zhao
,
2010b
:
Influence of the Asian–Pacific oscillation on spring precipitation over central eastern China
.
Adv. Atmos. Sci.
,
27
,
575
582
.
Zhou
,
B.
, and
P.
Zhao
,
2012
:
Simulating changes of spring Asian-Pacific Oscillation and associated atmospheric circulation in the mid-Holocene
.
Int. J. Climatol.
, 33, 529–538,
doi:10.1002/joc.3438
.
Zhou
,
B.
,
P.
Zhao
, and
X.
Cui
,
2010
:
Linkage between the Asian-Pacific Oscillation and the sea surface temperature in the North Pacific
.
Chin. Sci. Bull.
,
55
,
1193
1198
.
Zou
,
Y.
, and
P.
Zhao
,
2011
:
A study of the relationship between the Asian–Pacific oscillation and tropical cyclone activities over the coastal waters of China during autumn
.
Acta Meteor. Sin.
,
69
,
601
609
.